June 5, 1956 THE STEREOCHEMISTRY OF HYDRIDE REDUCTIONS

[CONTRIBUTION FROM THE CHEMICAL LABORATORY, UNIVERSITY OF CALIFORNIA]

2579

The Stereochemistry of Hydride Reductions BY WILLIAM G. DAUBEN, GERHARD J. FONKEN AND DONALD S. NOYCE

RECEIVED NOVEMBER 22, 1955

I t has been shown that reductions of alkylcyclohexanones with lithium aluminum hydride, sodium borohydride and alumi- num isopropoxide lead to increasing proportions of the axial (unstable) isomer. The mechanism of the reduction step is dis- cussed with reference to the ease of formation of the initial metallo-organic complex (steric approach control) and to the relative energetics of the formation of the products once the complex is formed (product development control).

The reduction by metal hydrides of ketones which can yield diastereomeric alcohols has been the sub- ject of numerous investigations and such studies have been carried out with a wide variety of ke- tones, including acyclic, alicyclic and polycyclic compounds.l Correlations of the stereochemical results have been proposed by numerous investiga-

and for the most part the results of these summaries have been expressed in terms of con- formational concepts, considering the most likely conformation of the initial ketone and the changes which must result during the reduction proper. One of the most descriptive postulates is that of Cram and Greene’ in which it was considered that the reduction by lithium aluminum hydride of acy- clic ketones proceeded by the formation of an ini- tial complex between the hydride and the carbonyl oxygen atom and in this complex, the oxygen “be- comes effectively the bulkiest group in the mole- cule and tends to orient itself between the two least bulky groups attached to the adjacent asymmetric carbon atom.” The predominant isomer “would be formed by approach of the entering group from the least hindered side of the carbonyl complex.”

Barton,Ic on the other hand, has merely summa- rized the results and has pointed out that with an unhindered ketone, the predominant product in the alicyclic series is the equatorial isomer. Other pre- vious workers, however, have paid little attention to the striking fact that the predominant product in any series usual ly is the more stable isomer. An excellent example of this generality is to be found in the study of the reduction of a series of 2-(a- piperazinyl) -benzylcyclohexanones8 in which many structural changes were made and in all cases the stable equatorial isomer was formed in higher yield.

In view of the foregoing considerations, two in- terpretations of the stereochemical results of the hydride reduction appear woithy of evaluation, one, a steric postulate involving competitive at- tacks from a favored (unhindered) or an unfavored (hindered) side which for simplicity can be termed “steric approach control” and the other, an energy

(1) For summaries, see (a) D. S. Noyce and D. B. Denney, THIS JOURNAL, 72, 5743 (1950); (b) D. J. Cram and F. A. Abd Elhafez. ibid., 74, 5828 (1952); (c) D. H. R. Barton, J. Chem. Sac., 1027 (1953).

(2) C. W. Shoppee and G H. R Summers, ibid., 687 (1950). (3) H. L. Goering and C. Serres, Jr., THIS JOURNAL, 74, 6908

(1952). (4) N. L. Paddock, Chemislry and Induslry , 63 (1953). (5) S. Bernstein, R. H. Lenhard and J . H. Williams, J . Org. Chem.,

(6) H. M. E. Cardwell, J. W. Cornfortb, S. R. Duff, H. Holtermann

(7) D. J . Cram and F. D. Greene, THIS JOURNAL, 76 , 6005 (1953). ( 8 ) P. B. Russell and R. Baltzly, ibid., 77, 620 (1955).

18, 1168 (1953).

and R. Robinson, J Chem. Soc., 361 (1953).

consideration involving the relative stability of the possible products and which can be termed “firoduct development control.” In the present in- vestigation, additional results with cyclic ketones have been obtained, using a variety of reduction methods, which yield further information of value in attempting to assess the various influences a t work in the control of the distribution of the iso- mers produced. These results are summarized in Table I.

TABIE I

DISTRIBUTION OF JSOMERS F‘JRMED BY REDUCTION OF

KETONES Y M e n t h o n e - 4-Methyl- 2-Methyl-

Neo- cyclo- cyclo- men- hexanone hexanon;

Menthol, thol, % % % % Reagent % Yo trans CIS trans CIS

LiAlH4 71 29’“ 81 19Ia 82 18 NaBH4 in metha-

no1 49 51 75 25 69 31 NaBH, in pyri-

dine . . . . BO 40 56 44 A1(O-Pri), 30 708 67 331° 42 58’” NaBH(OCH& . . . . . . . 70 30 Equilibrium Mostlyll 88 12 99 1

It is seen from the results in Table I that the re- duction of alkylcyclohexanones with lithium alumi-

.num hydride leads to the predominant formation of the stable isomer, while a change to sodium borohy- dride in either methanol or pyridine decreases the percentage of this isomer. I t should be particu- larly noted, however, that whereas the 2- and 4- methylcyclohexanones appear to yield the same proportion of isomers, the important feature to con- sider in evaluating the product development control is the difference of these ratios from the equilibrium mixture of isomers. An accurate determination of the position of the equilibrium for the two methyl- cyclohexanols was performed and the results also are listed in Table I. The position of equilibrium a t 99% trans for the 2-methylcyclohexanols as compared to 887, trans for the 4-methyl isomer is in line with conformational concepts. l2

On the basis of results obtained in this present study, a more nearly complete picture of the reduc- tion of the alkylcyclohexanones is a t hand and al- lows one, for the first time, to examine critically the

(9) 0. Zeitschel and H. Schmidt, Bcr., 69, 2303 (1926). (10) L. M. Jackman, A. K. Macbeth and J. A. Mills, J . Chcm.

(11) R. H. Pickard and W. 0. Littleburg, ibid., 101, 109 (1912). (12) For the case of the 2-methylcyclohexanols, previous workers

[R. Cornubert, M. Lafont-Lemoine and N. Nadjme-Ahadi, Comgl. rend., 837, 469 (1953)1, using the less precise phenylurethan method of analysis, have reported tha t the equilibrium mixture contains 95% t rans.

Sac., 2641 (1949).

2580 WILLIAX G. DAUBEX, GERHARD J. FONKEN AND DONALD S. NOYCE Vol. 78

effect of structure on isomer production. For the sake of simplicity, the case of menthone (I) will be considered first, for here the conformational ar- rangement of the original ketone is largely fixed with both alkyl groups in the equatorial arrange- ment. (To place both groups in an axial conforma-

I I1 111

tion would require an undue expenditure of en- ergy.) In this compound, the isopropyl group oc- cupies a position which flanks the carbonyl carbon very nearly equally on each side of the molecule thus making attack by a small reagent from either side equally favored as far as this substituent is concerned. The results of the lithium aluminum hydride reduction show that an axial approach resulting in the introduction of a hydrogen atom in an axial position and the formation of an equatorial hydroxyl group (11) (;.e., the formation of menthol) is the favored reaction. Thus, an example of mainly product development control is seen. LVhen menthone is reduced with sodium borohydride in methanol, a reagent which appears to be stereo- chemically more selective,' the amount of neomen- tho1 (111) (hydroxyl group axial) formed is larger, indicating that equatorial approach is becoming more important. This may be considered as indi- cating a greater effective size of the borohydride species and a case where steric approach control is beginning to exhibit competition with product de- velopment control. l3 The reduction of menthone with sodium borohydride in pyridine is compli- cated by the fact that concurrent with the reduc- tion of the carbonyl group, inversion of the asyni- metric center adjacent to the carbonyl group also occurs to some extent and hence no results of sig- nificance to this study are obtained.

In these reductions it is most probable that the reaction involves the preliminary formation of a metallo-organic complex between the carbonyl group and the hydride species. I t is in this com- plex that hydride transfer occurs and as the trans- fer begins, the sp2 bonding of the carbonyl group begins to change to the sp3 bonding of the product. In so doing, the intermediate stages of the reduction take on the steric requirements of the products and hence the relative stabilities of the products reflect themselves in this early stage of the reduc- tion. Thus, one sees there are two features of the reduction which must be examined, first, the ease of formation of the initial complex or steric approach control and, second, the relative energetics of the formation of the products once the complex is formed or product development control. In the case of the reduction of menthone with lithium alu-

(13) Various other workers have reported results which also indicate a greater effective size of the borohydride species. For example, T. G. Halsall, R. Hodges and E. R. H. Jones [I. C h r m . Soc., 3019 (1953)l have reported t h a t in methyl b-oxupolyporenate A a-acetate, lithium aluminum hydride reduces the carbonyl while sodium borohydride gives no reduction and H. B. Henbest, €3. R . H. Jones, A A Wagland and T. I. Wrigley [ i b i d . , 2477 (1955)], have reported tha t in a 9- bromo-1 1-oxo-steroid, lithium aluminum hydride attacks the hindered I 1 oxo group while sodium borohydride attacks only t h e q-hromo atom leaving t h e I Lcixo croti1) iintuucheil

minum hydride, since this reducing agent is a rela- tively small group and since, as discussed above, either side of the molecule ( i e . , an equatorial or axial attack) is equally favored sterically for attack of the hydride, the developing sp3 hybridization is essentially the controlling factor; the result is the formation of a greater predominance of the more stable isomer (equatorial hydroxyl). As the size of the reducing species is effectively increased, such as with sodium borohydride in methanol, considera- tion must be given to steric interactions between the P-axial hydrogen atoms on the ketone and the reducing moiety approaching in the axial position. This axial-axial type of interaction results in an increasing proportion of equatorial approach with the resultant formation of more of the less stable isomer with an axial hydroxyl group.I4

Interpretation of the results obtained with the 2- methyl- and the 4-methylcyclohexanones requires consideration of the two possible conformations of the ketones. I n the case of the 4-methyl isomer, due to the remote position of the alkyl substituent, the two conformations IV and V are identical as far as the substituents a and 6 to the carbonyl group are concerned.

In either conformation, approach of the reducing species from either side of the carbonyl group is equally favored and the reduction with lithium alu- minum hydride should be product development con- trolled and yield mainly an equatorial school. From conformation IV, this would result in the formation of mostly the trans isomer while from conformation V, a cis isomer would be formed. The relative con- tribution of each pathway will depend on the differ- ence in the energies of the various transition states. Since conformation V has an axial methyl group, i t would be expected to lead to transition states of higher energies than those from conformation IV. Consequently the former should be of lesser impor- tance but still allow for the formation of more cis isomer than from a compound with a frozen con- formation. It is seen in Table I that whereas the equilibrium mixture has only 12y0 of the cis, the lithium aluminum hydride reduction mixture con- tains 19% of the cis, a result in line with the predic- tion.

With the 2-methyl isomer, the presence of the methyl group on the carbon atom adjacent to the carbonyl group introduces further complexities. With regard to the approach of the reagent to the carbonyl group, in conformation 17 both equato- rial and axial approach is equally favored and reac- tion should be product development controlled. In conformation VII, the equatorial approach is sub- ject to some interference by the axial methyl

(14) I t should be et i iphasi~rd that the difference iii f r ee energy be tween any two isomers in such a series is small (-1-3 kcal.) and, in addition, the isomer ratio obtained is the result of an even smaller difference in free energy of activation (-0-1 kcal.) between the tw'o i l 1 : i c l i s Ic,i<linK to t h c t w < r Isomms

June 3 , 1936 THE STEREOCHEMISTRY OF HYDRIDE REDUCTIONS 258 1

VI1

group and reactions will be steric approach con- trolled. The former will yield a mixture of isomers similar to that found in the equilibrium mixture but the latter will yield a mixture more rich in the cis. Again, transition states arising from con- formation VI should be more favored, but the amount of cis isomer produced from the transition states from conformation VI1 should be more than in the 4-methyl case. Consequently, the total amount of cis isomer, as compared to the equilib- rium concentration, should be larger for the 2- methyl isomer when i t is reduced with lithium alu- minum hydride and it is seen from Table I that the equilibrium concentration of cis is 1% while 18% cis is formed in the reduction. Thus, the stable trans isomer predominates but the values differ from the equilibrium concentration in the proper direction.

Reduction of these niethylcyclohexanones with more bulky hydrides, such as sodium borohydride, is similar to that discussed with menthone. The result should be an increasing amount of the un- stable cis isomer and in the case of the 2-methyl compound this effect should be even more predomi- nant. These postulated results are borne out well in the results shown in Table I.

Many results obtained with different alicyclic compounds also are adequately explained by utiliza- tion of these same concepts. For example, in the reduction of cholestanone where the carbonyl group is unhindered by a- or &axial groups, the reaction should be product development controllei and it is found that the more stable equatorial iso- mer is formed in 90% yield by the use of lithium aluminum hydride. l6 When the more spacially demanding sodium borohydride is used, some steric approach control should be exerted and it is found that the amount of the equatorial isomer is less. 15c Reduction of an 11-ketosteroid illustrates the role of axial groups so placed that they may even block the axial approach of lithium aluminum hydride and allow the reaction to be completely steric approach controlled. In such a case, the prod- uct is the lip-axial hydroxyl c o m p o ~ n d . ~ * ~ ~

Experimental Reduction of Methylcyc1ohexanones.-The product com-

positions were determined by density measurements. The reference densities used for these analyses were those re- ported by Jackman, Macbeth and Mills.lo The d304 for the 4-methylcyclohexanols were: trans, 0.9040; cis, 0.9173. For the 2-methylcyclohexanols the values were: t rans , 0.9174; cis, 0.9274. The values for the trans isomers were in agreement with those obtained in this Laboratory for materials generated from the pure 3,5-dinitrobenzoate esters. All densities were taken with a calibrated 10-ml. pycnometer in a constant temperature bath calibrated against a National Bureau of Standards thermometer. _ _ _ _ ~

(15) (a) C. W. Shoppee and G. H. R . Summers, J . Chcm. Soc., 687 (1950); (b) H. R. Nace and G. I,. O'Connor, THIS JOURNAL, 73, 58% f lR31) , ( c ) W G Dauben, R A Slicheli and J. F. Eastham, i b i d , 74. 3853 11932).

a . Lithium Aluminum Hydride.-The reaction was con- ducted as described by Noyce and Denneyk and the prod- uct from 2-methylcyclohexanone had a d304 0,9191 (By0 cis) and a n m ~ 1.4625.

b. Sodium Borohydride in Methanol.-In a three- necked flask fitted with a sealed stirrer and a reflux conden- ser was placed 6.6 g. (0.178 mole) of sodium borohydride dissolved in 20 ml. of water which, in turn, had been diluted with 100 ml. of methanol. Over a period of 30 minutes, a solution of 20 g. (0.178 mole) of 4-methylcyclohexanone in 50 ml. of methanol was added at such a rate t o the stirred solution that a moderate reflux was obtained. After com- pletion of the addition, the mixture was refluxed for 4 hours, 150 ml. of 2 N sodium hydroxide added and the heating continued for an additional 2 hours.'6 The hydrolyzed solution was extracted continuously with ether for 24 hours and the product distilled through an 18" column, b.p. 90.- 91" (50 ".),I7 d304 0.9073 (25% cis), T Z * ~ D 1.4543, yield 18.3 g. (91%).

In a similar manner, 20 g. of 2-methylcyclohexanone was reduced to yield 18.2 g. (90%) of product, b.p. 90-92" (50

c. Sodium Borohydride in Pyridine.-The reduction was conducted as described in methanol except 6.16 g. of sodium borohydride was dissolved in 150 ml. of dry pyridine and the 20 g. of 4-methylcyclohexanone was dissolved in 5:) ml. of dry pyridine. After heating the reaction mixture at 70" for 6 hours, 100 ml. of 2 N sodium hydroxide was added and the heating continued for an additional 2 hours. Ex- traction of the reaction mixture with ether using a continu- ous extractor for 24 hours, followed by distillation yielded 14.3 g. (71%) of 4-methylcyclohexanol, b.p. 90-91" (50 mm.), dm4 0.9092 (40% cis), Y Z ~ ~ D 1.4566.

I n a similar manner, 20 g. of 2-methylcyclohexanone \vas reduced to yield 19.8 g. (97%) of product, b.p. 90-92" (50 mm.), d304 0.9218 (44y0 cis).

Sodium Trimethoxyborohydride in Pyridine .-In a 1-1. three-necked flask equipped. with a sealed stirrer, a thermometer and a dropping funnel, was placed 107 g. (0.6 mole) of sodium trimethoxyborohydride in 600 ml. of dry pyridine and then 60 g. (0.54 mole) of Z-methylcyclohexan- one was added, dropwise. The temperature rose to 47' upon the beginning of the addition and the rate of addition was adjusted to keep the temperature between 45-50', The addition required 45 minutes and then the reaction was stirred for an additional 3 hours at this same temperature. Upon cooling, the mixture was poured into 2-1. of water containing 200 ml. of concentrated hydrochloric acid and 500 g. of ice. After extraction with ether the product was distilled, b.p. 165-166', yield 36.8 g. (60%), d309 0.9204 (30% cis), T Z ~ O D 1.4627.

Equilibration of Methylcyclohexano1s.-The correspond- i i g alcohol (30 g.) was heated under reflux with 18.6 g. of aluminum isopropoxide and 1 ml. of acetone in 300 ml. of isopropyl alcohol for 72 hours. Time studies with each isomer showed that no change in isomer composition nc- curred after this reaction time. The reaction solution then was poured into 1500 ml. of water containing 100 ml. of con- centrated hydrochloric acid and the product extracted with ether.

For the 2-methyl isomer, the starting alcohol had a d30a

of 0.9218 (44y0 cis) and the final product a d304 of 0.9175 (1% cis). For the 4-methyl isomer the equilibrium was ap- proached from each side. In one case, the starting alcohol had a dma of 0.9094 f40% cis) and the final material a d30, of

mm.), d304 0.9205 (30% Cis), %"D 1.4625.

d .

0.9056 (12% cis) and & t h e 'second case, the starting afco- hol had a d304 of 0.9040 (0% cis) and the final product a d304

Reduction of I-Menthone with Sodium Borohvdride in of 0.9057 (12% cis).

Methanol.-The reduction was conducted as descri6ed above using 4.8 g. of sodium borohydride in 50 ml. of water and 150 ml. of methanol and 20 g. of 1-menthone ( [ c Y ] ~ ~ D -23.3, homog.) in 50 ml. of methanol. The product was distilled, b.p. 120-135" (55 mm.), yield 18.2 g., [ c Y ] ~ ~ D -15.9, homog.

(16) Low recovery of product was obtained with acid hydrolysis. It was ascertained t h a t t h e alkali treatment had no effect on the density of the product. I n a trial experiment the starting alcohol has a d 3 " r

of 0.9218 and t h e f i n d product after alkali treatment had a value of 0.9217.

(17) T h e absence of ketone in this and all subsequent products was shown by the absence of the carbonyl band in the infrared s])ectrum under such conditions where 2% would be detected.

2582 HERBERT C. BROWN AND B. C. SUBBA Kso Vol. 'is The product was reoxidized with sodium dichromate18 to menthone ( [ 0 1 ] * 5 ~ -23.7') showing no change in the con-

figuration of the starting material had occurred.

BERKELEY 4, CALIFORNIA (18) Organic Synlhcscs, Coll. Vol. I, John Wiley and Sons, Inc.,

New York 16, N. Y., 1944, p. 340.

[CONTRIBUTION FROM THE DEPARTMENT OF CHEMISTRY OF P U R D U E UNIVERSI? Y]

A New Powerful Reducing Agent-Sodium Borohydride in the Presence of Aluminum Chloride and Other Polyvalent Metal Halides1p2

BY HERBERT c. BROWX AND B. c. SUBB.4 RA03 RECEIVED NOVEMBER 11, 1955

The addition of aluminum chloride to sodium borohydride in the dimethyl ether of diethylene glycol (diglyme) gives a clear solution with enormously more powerful reducing properties than those of sodium borohydride itself. Esters, lactones and carboxylic acids are readily reduced to the corresponding alcohols. Sodium salts of the carboxylic acids are not re- duced, so that the reagent permits the selective reduction of an ester group in the presence of the carboxylate group. Nitro groups are not attacked by the reagent, permitting the ready reduction of nitro esters to the corresponding nitro alcohols. Nitriles are reduced t o primary amines in good yield. A survey was made of the reactivity of various functional groups toward the reagent by examining the hydrogen utilization by various representative compounds under standard reaction conditions. Simple acid amides form only salts (by reaction with one of the active hydrogen atoms of the amide group) which do not undergo reduction, whereas reduction does occur with the disub- stituted amides. Aldehydes, ketones, epoxides and anhydrides utilize hydride corresponding to reduction to the alcohol. Disulfides and sulfonyl chlorides are reduced, whereas a simple sulfone was not. Pyridine-N-oxide is reduced at 2 5 O , whereas pyridine itself does not react at this temperature. Benzyl chloride was not affected a t 25", but appeared t o undergo reac- tion at 75'. Other polyvalent metal halides. such as gallium trichloride and titanium tetrachloride, also appear capable of greatly increasing the reducing potentialities of sodium borohydride.

The results lead to the following conclusions.

In the past decade sodium b o r ~ h y d r i d e ~ , ~ and lithium aluminum hydrideav' have revolutionized the reduction of functional groups in organic chem- istry. Of these two reagents, sodium borohydride is a relatively mild reducing agent which is practi- cally specific for the carbonyl group in aldehydes and ketones. On the other hand, lithium alumi- num hydride is an exceedingly powerful reagent which attacks almost all reducible groups.

There would be advantages in possessing rea- gents with reducing capacities somewhere between these two extremes. One possible means of achiev- ing this end would be to increase the reducing prop- erties of the borohydride; another, to diminish the reducing properties of the aluminohydride. Both procedures are under active investigation in our laboratories.

It is quite clear that the reducing capacity of the borohydride is markedly affected by the metal ion present in the salt. Thus sodium borohydride re- duces typical esters, such as ethyl acetate and ethyl benzoate, only very slowly, whereas lithium boro-

reduces such esters quite easily.'O Kollonitsch, Fuchs and Gabor achieved the SUC-

cessful reduction of simple esters by sodium boro- hydride in tetrahydrofuran in the presence of lith-

(1) Addition Compounds of the Alkali Metal Halides. IV. ( 2 ) Based upon a thesis submitted by B C. Subba Rao in partial

fulfillment of the requirements for the degree of Doctor of Philosophy. (3) Parke, Davis and Co. Fellow a t Purdue University, 1952-1955. (4) H . I. Schlesinger, H. C. Brown, H . R . Hoekstra and L. R . Rapp,

( 5 ) S. W. Chaikin and W. G. Brown, ibid., 71, 122 (1949) ( G ) A. F. Finholt, A. C. Bond, Jr., and H . I. Schlesinger, ibid., 69,

( 7 ) R. F. Nystrotn and W. G. Brown, ibid., 69, 1197, 2548 (1947);

(8) €1. I . Schlesinger and H. C. Brown, ibid., 62, 3429 (1940). (9) R. F. Nystrom, S. W. Chaikin and W. G. Brown, ibid., 71, 3245

(1949). (IO) The ester groups in certain sugar lactones are readily reduced

even by sodium borohydride: M. L. Wolfrom and H . B. Wood, THIS JOURNAL, 73, 2933 (1951).

THIS JOURNAL, 76, 199 (1953).

3199 (1947).

70, 3738 (1948).

ium and magnesium iodides," as well as by the ap- plication of calcium, strontium and barium boro- hydrides. l2

The dimethyl ether of diethylene glycol (diglyme) is an excellent solvent for sodium borohydride. The addition of lithium bromide or magnesium chloride to such solutions provides a reagent which permits the ready reduction of esters.I3 In continuing this line of investigation, we have discovered that the addition of anhydrous aluminum chloride or other polyvalent metal halides, such as gallium chloride or titanium tetrachloride, provides solutions of tre- mendously enhanced reducing capacities. l 4 We, therefore, turned our attention to a detailed exami- nation of the potentialities of these new reducing systems.

Results Mixture of freshly prepared solutions of alumi-

num chloride ( 2 M ) and sodium borohydride (1.0 ill), in diglyme in a molar ratio AlClS: 3NaBH4 gives a clear solution which exhibits very powerful re- ducing properties. Preliminary experiments re- vealed that esters, such as ethyl benzoate, ethyl stearate and ethyl p-nitrobenzoate, were essentially completely reduced by the reagent in one hour a t room temperature.

The precise order of mixing the reagents does not appear to be important. A t least no significant dif- ference in results was observed in (1) adding the aluminum chloride solution to a mixture of SO- dium borohydride and ester in the solvent, (2) add- ing the sodium borohydride solution to a mixture of the ester and aluminum chloride in the solvent or (3) adding the ester to a mixture of aluminum chlo- ride and sodium borohydride in diglyme. We fi-

(11) J. Kollonitsch, P. Fuchs and V. Gabor, I V u I u ~ c , 173. 13.5 (1954) (12) J . Kollonitsch, P. Fuchs and V. Gabor, ibid., 175, 340 (1955). (13) H. C. Brown, E. J. Mead and B C. Subba Rao, T H I S JOURNAL.

(14) H. C. Brown and U. C. Siiblia Ran , & b i d . , 77, 3164 (195:J. 77, 6209 (195s).